Enterprise information integration must contextualise data served up to users from across multiple, disparate systems, writes Paul van Aswegen, GM of Informatica SA. Software-as-a-service (SaaS) can facilitate just that, with a raft of associated benefits.
Six reasons companies will take note are:
* It eliminates problems related to multiple operating systems;
* Minimises time and cost spent porting applications and data;
* Reduces time between releases;
* Removes the need to support legacy systems;
* Reduces the cost to SaaS providers; and
* Improves product testing.
SaaS relies on a multitenant architecture, which is its point of departure in providing benefits to data integration. Multitenant applications are the result of a combination of the application service provider model of the ‘90s and services such as Gmail that operate a single application instance serving many customers.
Databases have been rapidly adopted since the ‘60s. Today they proliferate throughout businesses. Additional complexity is introduced in groups of companies and merged entities. In many instances, individual business processes will run against their own database.
Silos make it difficult to access all data, reconcile data definitions and certify data quality.
But SaaS is no cure-all to the data integration ills that prevail in this diverse and complex environment. One of the first problems most organisations may encounter, when employing SaaS for data integration duty, is a number of different SaaS application programming interfaces (APIs). They may also encounter proprietary, incomplete or niche APIs.
With the number of applications underpinning distinct business processes present in most organisations, there will undoubtedly be a mix of vendors supplying them. It requires a healthy support for APIs from the data integration vendors. Another potential problem is the expertise and experience that data integration vendors have with SaaS applications.
But businesses are increasingly deploying SaaS applications, intent on getting the most from their heterogeneous and disparate data. The best approach is to integrate data through deploying SaaS applications supporting specific business processes.
That allows data integration hooks to be created at specific intervals in the business process, where they are most suitable to achieving the end result, which is enterprise information integration in support of operations.
Those intervals occur where the processes meet and dissect one another and data can be integrated at those points if the SaaS application APIs expose events.
Bearing in mind SaaS’s strengths and weaknesses, the right environment can harness its power to deliver against the data integration needs of businesses today.